Title¶

Developed by (Name: Rishitha Nimma)

Data Analysis¶

Importing the libraries necessary for your EDA¶

In [1]:
# Import your data easy to read
# Import your data easy to read
import numpy as np
import pandas as pd
import seaborn as sns
from scipy import stats
from sklearn.decomposition import PCA
import matplotlib.pyplot as plt
%matplotlib inline
import warnings
warnings.filterwarnings('ignore')
df = pd.read_excel('cardekho_dataset.xlsx')
df.head()
Out[1]:
brand model vehicle_age km_driven fuel_type transmission_type mileage engine seats selling_price
0 Maruti Alto 9 120000 Petrol Manual 19.70 796 5 120000
1 Hyundai Grand 5 20000 Petrol Manual 18.90 1197 5 550000
2 Hyundai i20 11 60000 Petrol Manual 17.00 1197 5 215000
3 Maruti Alto 9 37000 Petrol Manual 20.92 998 5 226000
4 Ford Ecosport 6 30000 Diesel Manual 22.77 1498 5 570000

Import libraries for EDA¶

In [2]:
# Import libraries for EDA 
# Import libraries for EDA 
## Imported required libraries in 2.1 already
# Import your data easy to read
import numpy as np
import pandas as pd
import seaborn as sns
from scipy import stats
import matplotlib.pyplot as plt
%matplotlib inline
import warnings
warnings.filterwarnings('ignore')

Review Data¶

In [3]:
# Check data dimension by using .shape
df.shape ## displays the shape of dataframe
Out[3]:
(15411, 10)
In [4]:
# Check types of data by .info() ; .dtype
df.info() # Displays DataFrame summary, including types. 

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 15411 entries, 0 to 15410
Data columns (total 10 columns):
 #   Column             Non-Null Count  Dtype  
---  ------             --------------  -----  
 0   brand              15411 non-null  object 
 1   model              15411 non-null  object 
 2   vehicle_age        15411 non-null  int64  
 3   km_driven          15411 non-null  int64  
 4   fuel_type          15411 non-null  object 
 5   transmission_type  15411 non-null  object 
 6   mileage            15411 non-null  float64
 7   engine             15411 non-null  int64  
 8   seats              15411 non-null  int64  
 9   selling_price      15411 non-null  int64  
dtypes: float64(1), int64(5), object(4)
memory usage: 1.2+ MB
In [5]:
df.dtypes  ## displays data types of all the variables in the dataframe
Out[5]:
brand                 object
model                 object
vehicle_age            int64
km_driven              int64
fuel_type             object
transmission_type     object
mileage              float64
engine                 int64
seats                  int64
selling_price          int64
dtype: object
In [6]:
# For quantitative variable, generate a table for the count, mean, standard deviation, minimum and maximum values and the quantities of the data 
df.describe() ## displays descriptive statistics summary for numeric variables
Out[6]:
vehicle_age km_driven mileage engine seats selling_price
count 15411.000000 1.541100e+04 15411.000000 15411.000000 15411.000000 1.541100e+04
mean 6.036338 5.561648e+04 19.701151 1486.057751 5.325482 7.749711e+05
std 3.013291 5.161855e+04 4.171265 521.106696 0.807628 8.941284e+05
min 0.000000 1.000000e+02 4.000000 793.000000 0.000000 4.000000e+04
25% 4.000000 3.000000e+04 17.000000 1197.000000 5.000000 3.850000e+05
50% 6.000000 5.000000e+04 19.670000 1248.000000 5.000000 5.560000e+05
75% 8.000000 7.000000e+04 22.700000 1582.000000 5.000000 8.250000e+05
max 29.000000 3.800000e+06 33.540000 6592.000000 9.000000 3.950000e+07

Generate Sub-dataset¶

In [7]:
# 1. Make sub-dataset(s) from your original dataset for your research objective, goals by dropping unnecessary variables 
# Creating sub-datasets for each of the three questions by dropping unnecessary variables

# Load the dataset again for subsetting purposes
data_sub = df.copy() # Creates a duplicate DataFrame copy.

# Sub-dataset for question 1: Factors contributing to high mileage
mileage_dataset = data_sub.drop(columns=['brand', 'model', 'selling_price','seats']) # Removes specified columns from DataFrame.

# Sub-dataset for question 2: Vehicle characteristics affecting price of automatic vs. manual transmission cars
transmission_dataset = data_sub.drop(columns=[ 'vehicle_age', 'km_driven', 'seats','brand','model'])

# Sub-dataset for question 3: Influence of fuel type on performance (mileage) and price
fuel_performance_dataset = data_sub.drop(columns=['brand', 'model', 'vehicle_age', 'km_driven','transmission_type', 'seats', 'engine'])

# Display the first few rows of each sub-dataset to ensure they are created correctly
mileage_dataset.head(), transmission_dataset.head(), fuel_performance_dataset.head()
Out[7]:
(   vehicle_age  km_driven fuel_type transmission_type  mileage  engine
 0            9     120000    Petrol            Manual    19.70     796
 1            5      20000    Petrol            Manual    18.90    1197
 2           11      60000    Petrol            Manual    17.00    1197
 3            9      37000    Petrol            Manual    20.92     998
 4            6      30000    Diesel            Manual    22.77    1498,
   fuel_type transmission_type  mileage  engine  selling_price
 0    Petrol            Manual    19.70     796         120000
 1    Petrol            Manual    18.90    1197         550000
 2    Petrol            Manual    17.00    1197         215000
 3    Petrol            Manual    20.92     998         226000
 4    Diesel            Manual    22.77    1498         570000,
   fuel_type  mileage  selling_price
 0    Petrol    19.70         120000
 1    Petrol    18.90         550000
 2    Petrol    17.00         215000
 3    Petrol    20.92         226000
 4    Diesel    22.77         570000)
In [8]:
# try to change variable names
# Renaming columns for better clarity in each sub-dataset

# Sub-dataset for question 1: Factors contributing to high mileage
mileage_dataset = mileage_dataset.rename(columns={
    'vehicle_age': 'age_of_vehicle',
    'transmission_type': 'gearbox_type',    
    'km_driven': 'kilometers_driven',
    'fuel_type': 'type_of_fuel',
    'mileage': 'car_mileage',
    'engine': 'engine_capacity'
}) # renaming the columns of dataframe

# Sub-dataset for question 2: Vehicle characteristics affecting price of automatic vs. manual transmission cars
transmission_dataset = transmission_dataset.rename(columns={
    'fuel_type': 'type_of_fuel',
    'transmission_type': 'gearbox_type',
    'mileage': 'car_mileage',
    'engine': 'engine_capacity',
    'selling_price': 'price'
})

# Sub-dataset for question 3: Influence of fuel type on performance (mileage) and price
fuel_performance_dataset = fuel_performance_dataset.rename(columns={
    'fuel_type': 'type_of_fuel',
    'mileage': 'car_mileage',
    'selling_price': 'price'
})

# Display the first few rows of each renamed sub-dataset
mileage_dataset.head(), transmission_dataset.head(), fuel_performance_dataset.head()
Out[8]:
(   age_of_vehicle  kilometers_driven type_of_fuel gearbox_type  car_mileage  \
 0               9             120000       Petrol       Manual        19.70   
 1               5              20000       Petrol       Manual        18.90   
 2              11              60000       Petrol       Manual        17.00   
 3               9              37000       Petrol       Manual        20.92   
 4               6              30000       Diesel       Manual        22.77   
 
    engine_capacity  
 0              796  
 1             1197  
 2             1197  
 3              998  
 4             1498  ,
   type_of_fuel gearbox_type  car_mileage  engine_capacity   price
 0       Petrol       Manual        19.70              796  120000
 1       Petrol       Manual        18.90             1197  550000
 2       Petrol       Manual        17.00             1197  215000
 3       Petrol       Manual        20.92              998  226000
 4       Diesel       Manual        22.77             1498  570000,
   type_of_fuel  car_mileage   price
 0       Petrol        19.70  120000
 1       Petrol        18.90  550000
 2       Petrol        17.00  215000
 3       Petrol        20.92  226000
 4       Diesel        22.77  570000)
In [9]:
# check if there are missing values
# Checking for missing values in all sub-datasets

# Checking the number of missing values in the main dataset
missing_values_main = df.isnull().sum()

# Checking missing values in each sub-dataset
missing_values_mileage = mileage_dataset.isnull().sum()
missing_values_transmission = transmission_dataset.isnull().sum()
missing_values_fuel_performance = fuel_performance_dataset.isnull().sum()

missing_values_main, missing_values_mileage, missing_values_transmission, missing_values_fuel_performance
Out[9]:
(brand                0
 model                0
 vehicle_age          0
 km_driven            0
 fuel_type            0
 transmission_type    0
 mileage              0
 engine               0
 seats                0
 selling_price        0
 dtype: int64,
 age_of_vehicle       0
 kilometers_driven    0
 type_of_fuel         0
 gearbox_type         0
 car_mileage          0
 engine_capacity      0
 dtype: int64,
 type_of_fuel       0
 gearbox_type       0
 car_mileage        0
 engine_capacity    0
 price              0
 dtype: int64,
 type_of_fuel    0
 car_mileage     0
 price           0
 dtype: int64)
In [10]:
# Check if there are duplicated values
# Checking for duplicates in the main dataset
duplicates_main = df.duplicated().sum()

# Checking for duplicates in each sub-dataset
duplicates_mileage = mileage_dataset.duplicated().sum()
duplicates_transmission = transmission_dataset.duplicated().sum()
duplicates_fuel_performance = fuel_performance_dataset.duplicated().sum()

duplicates_main, duplicates_mileage, duplicates_transmission, duplicates_fuel_performance
Out[10]:
(189, 1919, 5858, 6391)

Checking Outliers¶

In [11]:
# check if quantitative variables have outliers.
# Check for outliers using the IQR method for each numeric variable

def detect_outliers(dataset, numeric_columns):
    outlier_report = {}
    for column in numeric_columns:
        Q1 = dataset[column].quantile(0.25)
        Q3 = dataset[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        outliers = ((dataset[column] < lower_bound) | (dataset[column] > upper_bound)).sum()
        outlier_report[column] = outliers
    return outlier_report

# Detecting outliers in each sub-dataset

numeric_columns_mileage = ['age_of_vehicle', 'kilometers_driven', 'car_mileage', 'engine_capacity']
numeric_columns_transmission = ['car_mileage', 'engine_capacity', 'price']
numeric_columns_fuel_performance = ['car_mileage', 'price']

# Outliers for mileage dataset
outliers_mileage = detect_outliers(mileage_dataset, numeric_columns_mileage)

# Outliers for transmission dataset
outliers_transmission = detect_outliers(transmission_dataset, numeric_columns_transmission)

# Outliers for fuel performance dataset
outliers_fuel_performance = detect_outliers(fuel_performance_dataset, numeric_columns_fuel_performance)

outliers_mileage, outliers_transmission, outliers_fuel_performance
Out[11]:
({'age_of_vehicle': 154,
  'kilometers_driven': 466,
  'car_mileage': 88,
  'engine_capacity': 2130},
 {'car_mileage': 88, 'engine_capacity': 2130, 'price': 1386},
 {'car_mileage': 88, 'price': 1386})
In [12]:
# Function to remove outliers based on IQR method
def remove_outliers(dataset, numeric_columns):
    for column in numeric_columns:
        Q1 = dataset[column].quantile(0.25)
        Q3 = dataset[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        dataset = dataset[(dataset[column] >= lower_bound) & (dataset[column] <= upper_bound)]
    return dataset


# Remove outliers for mileage dataset
mileage_dataset_clean = remove_outliers(mileage_dataset, numeric_columns_mileage)

# Remove outliers for transmission dataset
transmission_dataset_clean = remove_outliers(transmission_dataset, numeric_columns_transmission)

# Remove outliers for fuel performance dataset
fuel_performance_dataset_clean = remove_outliers(fuel_performance_dataset, numeric_columns_fuel_performance)

# Display the cleaned datasets
mileage_dataset_clean.head(), transmission_dataset_clean.head(), fuel_performance_dataset_clean.head()
Out[12]:
(   age_of_vehicle  kilometers_driven type_of_fuel gearbox_type  car_mileage  \
 0               9             120000       Petrol       Manual        19.70   
 1               5              20000       Petrol       Manual        18.90   
 2              11              60000       Petrol       Manual        17.00   
 3               9              37000       Petrol       Manual        20.92   
 4               6              30000       Diesel       Manual        22.77   
 
    engine_capacity  
 0              796  
 1             1197  
 2             1197  
 3              998  
 4             1498  ,
   type_of_fuel gearbox_type  car_mileage  engine_capacity   price
 0       Petrol       Manual        19.70              796  120000
 1       Petrol       Manual        18.90             1197  550000
 2       Petrol       Manual        17.00             1197  215000
 3       Petrol       Manual        20.92              998  226000
 4       Diesel       Manual        22.77             1498  570000,
   type_of_fuel  car_mileage   price
 0       Petrol        19.70  120000
 1       Petrol        18.90  550000
 2       Petrol        17.00  215000
 3       Petrol        20.92  226000
 4       Diesel        22.77  570000)

Generating Plot(s)¶

In [13]:
# generate plots to support your objective and goals
import matplotlib.pyplot as plt
import seaborn as sns

# Set up Seaborn style
sns.set(style="whitegrid")

# Box Plot for factors contributing to high mileage
plt.figure(figsize=(16, 12))

# Box plot for Age of Vehicle
plt.subplot(2, 2, 1)
sns.boxplot(x='age_of_vehicle', y='car_mileage', data=mileage_dataset_clean)
plt.title('Mileage by Age of Vehicle')
plt.xlabel('Age of Vehicle')
plt.ylabel('Mileage (kmpl)')

# Box plot for Kilometers Driven
plt.subplot(2, 2, 2)
sns.boxplot(x='gearbox_type', y='car_mileage', data=mileage_dataset_clean)
plt.title('Mileage by Kilometers Driven')
plt.xlabel('Kilometers Driven')
plt.ylabel('Mileage (kmpl)')

# Box plot for Fuel Type
plt.subplot(2, 2, 3)
sns.boxplot(x='type_of_fuel', y='car_mileage', data=mileage_dataset_clean)
plt.title('Mileage by Fuel Type')
plt.xlabel('Fuel Type')
plt.ylabel('Mileage (kmpl)')

# Box plot for Engine Capacity
plt.subplot(2, 2, 4)
sns.boxplot(x='engine_capacity', y='car_mileage', data=mileage_dataset_clean)
plt.title('Mileage by Engine Capacity')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Mileage (kmpl)')

plt.tight_layout()
plt.show()
No description has been provided for this image

The box plots above provide insights into the factors contributing to high mileage.

  • Older vehicles tend to have slightly lower mileage, but the variation is not extreme.
  • Manual transmission vehicles tend to offer a higher mileage compared to automatic transmission vehicles, which shows a tighter distribution.
  • Diesel vehicles generally offer higher mileage compared to petrol vehicles, as reflected in the higher median mileage for diesel cars.
  • Smaller engine capacities tend to have higher mileage, while larger engines show lower mileage values, which is expected due to higher fuel consumption.
In [14]:
import matplotlib.pyplot as plt
import seaborn as sns

# Set up Seaborn style
sns.set(style="whitegrid")

# Histogram for Kilometers Driven (Distribution)
plt.figure(figsize=(10, 6))
sns.histplot(mileage_dataset_clean['kilometers_driven'], kde=True, color='blue', bins=20)
plt.title('Distribution of Kilometers Driven')
plt.xlabel('Kilometers Driven')
plt.ylabel('Frequency')
plt.show()

# Histogram for Engine Capacity (Distribution)
plt.figure(figsize=(10, 6))
sns.histplot(mileage_dataset_clean['engine_capacity'], kde=True, color='green', bins=20)
plt.title('Distribution of Engine Capacity')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
No description has been provided for this image

The histograms above provide a clear view of the distribution of two key variables in the dataset:

  • The distribution shows that most vehicles have driven a moderate amount of kilometers, with fewer cars at both the lower and higher ends. The distribution is slightly skewed, indicating that there are more cars with higher kilometers driven.
  • This distribution highlights that the majority of the vehicles have engine capacities clustered around certain common values (such as 1000cc to 1500cc), with fewer vehicles having engines that are either much smaller or much larger.
In [15]:
# Scatter Plot for Kilometers Driven vs Car Mileage
plt.figure(figsize=(10, 6))
sns.scatterplot(x='engine_capacity', y='car_mileage', data=transmission_dataset_clean)
plt.title('Engine Cpacity vs. Car Mileage')
plt.xlabel('Engine Cpacity')
plt.ylabel('Car Mileage (kmpl)')
plt.show()
No description has been provided for this image

The scatter plot above shows the relationship between engine capacity and car mileage. Here are some key insights:

  • As the engine capacity increases, there is a noticeable decline in car mileage, which is expected as larger engines typically consume more fuel.
In [16]:
#  Hexbin Plot for Kilometers Driven vs Price (Alternative to Scatter Plot)
plt.figure(figsize=(10, 6))
plt.hexbin(fuel_performance_dataset_clean['price'], fuel_performance_dataset_clean['car_mileage'], gridsize=30, cmap='Blues')
plt.title('Hexbin Plot of Price vs. Car Mileage')
plt.xlabel('Price')
plt.ylabel('Car Mileage (kmpl)')
plt.colorbar(label='Density')
plt.show()
No description has been provided for this image
  • The darker hexagons represent areas where many vehicles cluster around a particular price and mileage. These clusters indicate typical pricing ranges for vehicles with certain mileage.
  • There is a slight tendency for higher-priced vehicles to have lower mileage, although the relationship isn't overly strong, suggesting other factors likely influence price beyond just mileage.

EDA¶

Question 1: What factors contribute to high mileage in used cars ?

In [17]:
mileage_dataset_clean.describe()
Out[17]:
age_of_vehicle kilometers_driven car_mileage engine_capacity
count 12869.000000 12869.000000 12869.000000 12869.000000
mean 5.789339 48798.379750 20.596419 1319.224959
std 2.810510 27197.863832 3.521140 284.008205
min 0.000000 100.000000 10.000000 793.000000
25% 4.000000 27700.000000 18.100000 1197.000000
50% 5.000000 47592.000000 20.360000 1248.000000
75% 8.000000 67000.000000 23.000000 1497.000000
max 14.000000 130000.000000 30.480000 2157.000000
In [18]:
# perform your EDA
# Univariate Analysis

import matplotlib.pyplot as plt
import seaborn as sns

# Set up Seaborn style
sns.set(style="whitegrid")

# Summary statistics for Age of Vehicle
print("Summary Statistics for Age of Vehicle:")
print(mileage_dataset_clean['age_of_vehicle'].describe())

# Summary statistics for Kilometers Driven
print("\nSummary Statistics for Kilometers Driven:")
print(mileage_dataset_clean['kilometers_driven'].describe())

# Frequency count for Fuel Type
print("\nFuel Type Distribution:")
print(mileage_dataset_clean['type_of_fuel'].value_counts())

# Summary statistics for Engine Capacity
print("\nSummary Statistics for Engine Capacity:")
print(mileage_dataset_clean['engine_capacity'].describe())

# Summary statistics for Mileage
print("\nSummary Statistics for Mileage:")
print(mileage_dataset_clean['car_mileage'].describe())

# Frequency count for Fuel Type
print("\nTransmission Type Distribution:")
print(mileage_dataset_clean['gearbox_type'].value_counts())

# Visualizations
plt.figure(figsize=(16, 12))

Summary Statistics for Age of Vehicle:
count    12869.000000
mean         5.789339
std          2.810510
min          0.000000
25%          4.000000
50%          5.000000
75%          8.000000
max         14.000000
Name: age_of_vehicle, dtype: float64

Summary Statistics for Kilometers Driven:
count     12869.000000
mean      48798.379750
std       27197.863832
min         100.000000
25%       27700.000000
50%       47592.000000
75%       67000.000000
max      130000.000000
Name: kilometers_driven, dtype: float64

Fuel Type Distribution:
Petrol    7331
Diesel    5269
CNG        227
LPG         42
Name: type_of_fuel, dtype: int64

Summary Statistics for Engine Capacity:
count    12869.000000
mean      1319.224959
std        284.008205
min        793.000000
25%       1197.000000
50%       1248.000000
75%       1497.000000
max       2157.000000
Name: engine_capacity, dtype: float64

Summary Statistics for Mileage:
count    12869.000000
mean        20.596419
std          3.521140
min         10.000000
25%         18.100000
50%         20.360000
75%         23.000000
max         30.480000
Name: car_mileage, dtype: float64

Transmission Type Distribution:
Manual       10437
Automatic     2432
Name: gearbox_type, dtype: int64
Out[18]:
<Figure size 1600x1200 with 0 Axes>
<Figure size 1600x1200 with 0 Axes>
In [19]:
# Histogram for Age of Vehicle
plt.figure(figsize=(12, 8))
plt.subplot(2, 2, 1)
sns.histplot(mileage_dataset_clean['age_of_vehicle'], kde=True, bins=20, color='blue')
plt.title('Distribution of Age of Vehicle')
plt.xlabel('Age of Vehicle (years)')
plt.ylabel('Frequency')

# Histogram for Kilometers Driven
plt.subplot(2, 2, 2)
sns.histplot(mileage_dataset_clean['kilometers_driven'], kde=True, bins=20, color='green')
plt.title('Distribution of Kilometers Driven')
plt.xlabel('Kilometers Driven')
plt.ylabel('Frequency')

# Bar plot for Fuel Type
plt.subplot(2, 2, 3)
sns.countplot(x='type_of_fuel', data=mileage_dataset_clean, palette='Set2')
plt.title('Fuel Type Distribution')
plt.xlabel('Fuel Type')
plt.ylabel('Count')

# Histogram for Engine Capacity
plt.subplot(2, 2, 4)
sns.histplot(mileage_dataset_clean['engine_capacity'], kde=True, bins=20, color='purple')
plt.title('Distribution of Engine Capacity')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Frequency')

plt.tight_layout()
plt.show()
No description has been provided for this image
In [20]:
# Histogram for Mileage
plt.figure(figsize=(10, 6))
sns.histplot(mileage_dataset_clean['car_mileage'], kde=True, bins=20, color='red')
plt.title('Distribution of Car Mileage')
plt.xlabel('Mileage (kmpl)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [21]:
# Bar plot for Fuel Type
plt.figure(figsize=(10, 6))
sns.countplot(x='gearbox_type', data=mileage_dataset_clean, palette='Set2')
plt.title('Transmission Type Distribution')
plt.xlabel('Transmission Type')
plt.ylabel('Count')
Out[21]:
Text(0, 0.5, 'Count')
No description has been provided for this image
In [22]:
# Bivariate Analysis
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np

# Set up Seaborn style
sns.set(style="whitegrid")

# Relationship between Age of Vehicle and Car Mileage
plt.figure(figsize=(10, 6))
sns.scatterplot(x='age_of_vehicle', y='car_mileage', data=mileage_dataset_clean, color='blue')
plt.title('Age of Vehicle vs. Car Mileage')
plt.xlabel('Age of Vehicle (years)')
plt.ylabel('Car Mileage (kmpl)')
plt.show()

# Calculate correlation
correlation_age_mileage = np.corrcoef(mileage_dataset_clean['age_of_vehicle'], mileage_dataset_clean['car_mileage'])[0, 1]
print(f"Correlation between Age of Vehicle and Car Mileage: {correlation_age_mileage:.2f}")
No description has been provided for this image 
Correlation between Age of Vehicle and Car Mileage: -0.23
In [23]:
# Relationship between Kilometers Driven and Car Mileage
plt.figure(figsize=(10, 6))
sns.scatterplot(x='kilometers_driven', y='car_mileage', data=mileage_dataset_clean, color='green')
plt.title('Kilometers Driven vs. Car Mileage')
plt.xlabel('Kilometers Driven')
plt.ylabel('Car Mileage (kmpl)')
plt.show()

# Calculate correlation
correlation_km_mileage = np.corrcoef(mileage_dataset_clean['kilometers_driven'], mileage_dataset_clean['car_mileage'])[0, 1]
print(f"Correlation between Kilometers Driven and Car Mileage: {correlation_km_mileage:.2f}")
No description has been provided for this image 
Correlation between Kilometers Driven and Car Mileage: 0.06
In [24]:
# Comparison of Mileage by Fuel Type
plt.figure(figsize=(10, 6))
sns.boxplot(x='type_of_fuel', y='car_mileage', data=mileage_dataset_clean, palette='Set2')
plt.title('Mileage by Fuel Type')
plt.xlabel('Fuel Type')
plt.ylabel('Car Mileage (kmpl)')
plt.show()

# Summary statistics for mileage by fuel type
print("\nSummary Statistics for Mileage by Fuel Type:")
print(mileage_dataset_clean.groupby('type_of_fuel')['car_mileage'].describe())
No description has been provided for this image 
Summary Statistics for Mileage by Fuel Type:
               count       mean       std    min     25%    50%    75%    max
type_of_fuel                                                                 
CNG            227.0  23.703392  4.569385  11.88  20.880  26.20  26.60  30.48
Diesel        5269.0  22.305422  3.617784  10.00  19.870  22.54  25.20  28.40
LPG             42.0  17.953571  3.805650  13.45  15.275  17.30  18.70  26.20
Petrol        7331.0  19.287045  2.735209  10.50  17.400  18.90  21.01  28.09
In [25]:
# Relationship between Engine Capacity and Car Mileage
plt.figure(figsize=(10, 6))
sns.scatterplot(x='engine_capacity', y='car_mileage', data=mileage_dataset_clean, color='purple')
plt.title('Engine Capacity vs. Car Mileage')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Car Mileage (kmpl)')
plt.show()

# Calculate correlation
correlation_engine_mileage = np.corrcoef(mileage_dataset_clean['engine_capacity'], mileage_dataset_clean['car_mileage'])[0, 1]
print(f"Correlation between Engine Capacity and Car Mileage: {correlation_engine_mileage:.2f}")
No description has been provided for this image 
Correlation between Engine Capacity and Car Mileage: -0.38
In [26]:
# Set up Seaborn style
sns.set(style="whitegrid")

# Pair plot for multiple relationships, including Age of Vehicle and Car Mileage
plt.figure(figsize=(12, 10))
# Generate the pairplot
pairplot = sns.pairplot(mileage_dataset_clean[['age_of_vehicle', 'kilometers_driven', 'car_mileage', 'engine_capacity']], diag_kind='kde')

# Set the title outside the plot using plt.suptitle
plt.suptitle('Pairplot of Vehicle Age, Kilometers Driven, Car Mileage, and Engine Capacity', 
             y=1.02, fontsize=16)

# Adjust the layout to prevent title overlap
plt.tight_layout()

plt.show()

<Figure size 1200x1000 with 0 Axes>
No description has been provided for this image
In [27]:
#  Bar Plot of Mean Mileage by Fuel Type
plt.figure(figsize=(10, 6))
fuel_mileage_mean = mileage_dataset_clean.groupby('type_of_fuel')['car_mileage'].mean().reset_index()
sns.barplot(x='type_of_fuel', y='car_mileage', data=fuel_mileage_mean, palette='Set2')
plt.title('Mean Mileage by Fuel Type')
plt.xlabel('Fuel Type')
plt.ylabel('Mean Car Mileage (kmpl)')
plt.show()
No description has been provided for this image
In [28]:
# multivariate analysis

# Calculate the correlation matrix
correlation_matrix = mileage_dataset_clean[['age_of_vehicle', 'kilometers_driven', 'engine_capacity', 'car_mileage']].corr()

# Create a heatmap
plt.figure(figsize=(10, 6))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', linewidths=0.5)
plt.title('Heatmap of Correlations Between Variables')
plt.show()
No description has been provided for this image
In [29]:
# Create a FacetGrid to examine the relationship between Age of Vehicle, Kilometers Driven, and Mileage by Fuel Type
g = sns.FacetGrid(mileage_dataset_clean, col="type_of_fuel", height=5, aspect=1)
g.map(sns.scatterplot, "age_of_vehicle", "car_mileage", alpha=0.6)
g.add_legend()

# Add titles and labels
g.set_axis_labels("Age of Vehicle (years)", "Car Mileage (kmpl)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('FacetGrid: Mileage vs. Age of Vehicle by Fuel Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [30]:
# FacetGrid for Kilometers Driven vs. Car Mileage by Fuel Type
g = sns.FacetGrid(mileage_dataset_clean, col="type_of_fuel", height=5, aspect=1)
g.map(sns.scatterplot, "kilometers_driven", "car_mileage", alpha=0.6)
g.add_legend()

# Add titles and labels
g.set_axis_labels("Kilometers Driven", "Car Mileage (kmpl)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('FacetGrid: Mileage vs. Kilometers Driven by Fuel Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [31]:
# FacetGrid for Engine Capacity vs. Car Mileage by Fuel Type
g = sns.FacetGrid(mileage_dataset_clean, col="type_of_fuel", height=5, aspect=1)
g.map(sns.scatterplot, "engine_capacity", "car_mileage", alpha=0.6)
g.add_legend()

# Add titles and labels
g.set_axis_labels("Engine Capacity (cc)", "Car Mileage (kmpl)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('FacetGrid: Mileage vs. Engine Capacity by Fuel Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [32]:
# FacetGrid to display transmission type vs car mileage faceted by fuel type
g = sns.FacetGrid(mileage_dataset_clean, col="type_of_fuel", height=5, aspect=1)
g.map(sns.stripplot, "gearbox_type", "car_mileage", jitter=True, palette="Set2", alpha=0.7)
g.add_legend()

# Add titles and labels
g.set_axis_labels("Transmission Type", "Car Mileage (kmpl)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('Stripplot: Transmission Type vs Car Mileage Faceted by Fuel Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image

Question 2: How does the combination of vehicle characteristics affect the selling price of automatic vs. manual transmission cars?

In [33]:
transmission_dataset_clean.describe()
Out[33]:
car_mileage engine_capacity price
count 12395.000000 12395.000000 1.239500e+04
mean 20.788276 1278.929407 5.371584e+05
std 3.458984 241.935832 2.418689e+05
min 10.000000 793.000000 4.000000e+04
25% 18.500000 1197.000000 3.500000e+05
50% 20.510000 1248.000000 5.000000e+05
75% 23.010000 1497.000000 6.800000e+05
max 30.480000 2148.000000 1.280000e+06
In [34]:
# Univariate Analysis
import matplotlib.pyplot as plt

# Pie chart for Transmission Type
plt.figure(figsize=(8, 6))
transmission_counts = transmission_dataset_clean['gearbox_type'].value_counts()
plt.pie(transmission_counts, labels=transmission_counts.index, autopct='%1.1f%%', startangle=90, colors=['#ff9999','#66b3ff'])
plt.title('Pie Chart: Transmission Type Distribution')
plt.show()

# Pie chart for Fuel Type
plt.figure(figsize=(8, 6))
fuel_counts = transmission_dataset_clean['type_of_fuel'].value_counts()

# Create pie chart without labels in the pie, and add a legend instead
plt.pie(fuel_counts, autopct='%1.1f%%', startangle=90, colors=['#ff9999','#66b3ff', '#99ff99', '#ffcc99'])
plt.title('Pie Chart: Fuel Type Distribution')

# Add a legend to show the labels
plt.legend(labels=fuel_counts.index, loc="best")
plt.tight_layout()
plt.show()
No description has been provided for this image
No description has been provided for this image
In [35]:
plt.figure(figsize=(8, 6))
sns.histplot(transmission_dataset_clean['engine_capacity'], kde=True, bins=5, color='green')
plt.title('Histogram: Engine Capacity Distribution')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [36]:
# Density Plot for Selling Price
plt.figure(figsize=(8, 6))
sns.kdeplot(transmission_dataset_clean['price'], fill=True, color='purple')
plt.title('Density Plot: Selling Price Distribution')
plt.xlabel('Price (INR)')
plt.ylabel('Density')
plt.show()
No description has been provided for this image
In [37]:
#  Density Plot for Selling Price
plt.figure(figsize=(8, 6))
sns.kdeplot(transmission_dataset_clean['car_mileage'], fill=True, color='purple')
plt.title('Density Plot: Car Mileage Distribution')
plt.xlabel('Car Mileage (Km)')
plt.ylabel('Density')
plt.show()
No description has been provided for this image
In [38]:
# Bivariate Analysis
# Histogram for Selling Price
plt.figure(figsize=(10, 6))
sns.histplot(data=transmission_dataset_clean, x='price', hue='gearbox_type', kde=True, bins=30, palette="Set2")
plt.title('Histogram: Selling Price Distribution by Transmission Type')
plt.xlabel('Selling Price (INR)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [39]:
# Histogram for Engine Capacity
plt.figure(figsize=(10, 6))
sns.histplot(data=transmission_dataset_clean, x='engine_capacity', hue='gearbox_type', kde=True, bins=30, palette="Set2")
plt.title('Histogram: Engine Capacity Distribution by Transmission Type')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [40]:
# Scatter Plot for Car Mileage vs. Price with Gearbox Type as Hue
plt.figure(figsize=(10, 6))
sns.scatterplot(x='car_mileage', y='price',  data=transmission_dataset_clean, palette='Set2', s=100)
plt.title('Scatter Plot: Car Mileage vs. Price ')
plt.xlabel('Car Mileage (kmpl)')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [41]:
# Scatter Plot for Car Mileage vs. Price with Gearbox Type as Hue
plt.figure(figsize=(10, 6))
sns.scatterplot(x='gearbox_type', y='car_mileage',  data=transmission_dataset_clean, palette='Set2', s=100)
plt.title('Scatter Plot: GearBox Type vs. Car Mileage')
plt.xlabel('GearBox Type')
plt.ylabel('Car Mileage')
plt.show()
No description has been provided for this image
In [42]:
# Box Plot for Type of Fuel vs. Price by Gearbox Type
plt.figure(figsize=(10, 6))
sns.boxplot(x='type_of_fuel', y='price', data=transmission_dataset_clean, palette='Set2')
plt.title('Box Plot: Fuel Type vs. Price ')
plt.xlabel('Fuel Type')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [43]:
# Box Plot for Type of Fuel vs. Price by Gearbox Type
plt.figure(figsize=(10, 6))
sns.scatterplot(x='gearbox_type', y='type_of_fuel', data=transmission_dataset_clean, palette='Set2')
plt.title('Box Plot: Gearbox Type  vs. Fuel Type ')
plt.xlabel('Gearbox Type')
plt.ylabel('Fuel Type')
plt.show()
No description has been provided for this image
In [44]:
# Scatter Plot for Engine Capacity vs. Price with Gearbox Type as Hue
plt.figure(figsize=(10, 6))
sns.scatterplot(x='engine_capacity', y='price', data=transmission_dataset_clean, palette='Set2', s=100)
plt.title('Scatter Plot: Engine Capacity vs. Price ')
plt.xlabel('Engine Capacity (cc)')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [45]:
# Scatter Plot for Engine Capacity vs. Price with Gearbox Type as Hue
plt.figure(figsize=(10, 6))
sns.scatterplot(x='gearbox_type', y='engine_capacity', data=transmission_dataset_clean, palette='Set2')
plt.title('Scatter Plot: Gearbox Type vs. Engine Capacity ')
plt.xlabel('Gearbox Type')
plt.ylabel('Engine Capacity')
plt.show()
No description has been provided for this image
In [46]:
# Violin Plot for Type of Fuel vs. Price by Gearbox Type
plt.figure(figsize=(10, 6))
sns.violinplot(x='type_of_fuel', y='price', data=transmission_dataset_clean, split=True, palette='Set2')
plt.title('Violin Plot: Fuel Type vs. Price ')
plt.xlabel('Fuel Type')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [47]:
plt.figure(figsize=(10, 6))
sns.countplot(x='gearbox_type', hue='type_of_fuel', data=transmission_dataset_clean, palette='Set2')
plt.title('Count Plot: Gearbox Type by Fuel Type')
plt.xlabel('Gearbox Type')
plt.ylabel('Count')
plt.show()
No description has been provided for this image
In [48]:
# Multivariate Analysis
#  Pair Plot for Car Mileage, Engine Capacity, and Price with Gearbox Type as Hue
sns.pairplot(transmission_dataset_clean, hue='gearbox_type', diag_kind='kde', palette='Set2', height=2.5)
plt.suptitle('Pair Plot: Car Mileage, Engine Capacity, and Price by Gearbox Type', y=1.02)
plt.show()
No description has been provided for this image
In [49]:
transmission_dataset_clean.head()
Out[49]:
type_of_fuel gearbox_type car_mileage engine_capacity price
0 Petrol Manual 19.70 796 120000
1 Petrol Manual 18.90 1197 550000
2 Petrol Manual 17.00 1197 215000
3 Petrol Manual 20.92 998 226000
4 Diesel Manual 22.77 1498 570000
In [50]:
# Plotting type_of_fuel vs price with hue as gearbox_type

plt.figure(figsize=(10, 6))
sns.boxplot(x='type_of_fuel', y='price', hue='gearbox_type', data=transmission_dataset_clean, palette='Set2')
plt.title('Box Plot: Fuel Type vs. Price by Gearbox Type')
plt.xlabel('Fuel Type')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [51]:
# FacetGrid: Car Mileage vs Price by Fuel Type and Gearbox Type
g = sns.FacetGrid(transmission_dataset_clean, col="type_of_fuel", hue="gearbox_type", height=5, aspect=1.2)
g.map(sns.scatterplot, "car_mileage", "price", s=100, alpha=0.7)
g.add_legend()
g.set_axis_labels("Car Mileage (kmpl)", "Price (INR)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('FacetGrid: Price vs. Car Mileage by Fuel Type and Gearbox Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [52]:
# Heatmap: Correlation of Continuous Variables
corr_matrix = transmission_dataset_clean[['car_mileage', 'engine_capacity', 'price']].corr()
plt.figure(figsize=(8, 6))
sns.heatmap(corr_matrix, annot=True, cmap="coolwarm", linewidths=0.5)
plt.title('Heatmap of Correlation Matrix')
plt.show()
No description has been provided for this image

Question 3 : How does fuel type (Petrol vs. Diesel) influence car performance in terms of mileage and price?

In [53]:
fuel_performance_dataset_clean.describe()
Out[53]:
car_mileage price
count 13948.000000 1.394800e+04
mean 20.072265 5.747825e+05
std 3.899249 2.785681e+05
min 9.000000 4.000000e+04
25% 17.400000 3.700000e+05
50% 20.000000 5.300000e+05
75% 22.770000 7.250000e+05
max 30.480000 1.485000e+06
In [54]:
# Univariate Analysis
# Pie chart for Fuel Type
plt.figure(figsize=(8, 6))
fuel_counts = fuel_performance_dataset_clean['type_of_fuel'].value_counts()

# Create pie chart without labels in the pie, and add a legend instead
plt.pie(fuel_counts, autopct='%1.1f%%', startangle=90, colors=['#ff9999','#66b3ff', '#99ff99', '#ffcc99'])
plt.title('Pie Chart: Fuel Type Distribution')

# Add a legend to show the labels
plt.legend(labels=fuel_counts.index, loc="best")
plt.tight_layout()
plt.show()
No description has been provided for this image
In [55]:
# Histogram for Car Mileage
plt.figure(figsize=(8, 6))
sns.histplot(fuel_performance_dataset_clean['car_mileage'], kde=True, bins=5, color='green')
plt.title('Histogram: Car Mileage Distribution')
plt.xlabel('Car Mileage (kmpl)')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [56]:
#  Density Plot for Price Distribution
plt.figure(figsize=(8, 6))
sns.kdeplot(fuel_performance_dataset_clean['price'], fill=True, color='purple')
plt.title('Density Plot: Price Distribution')
plt.xlabel('Price (INR)')
plt.ylabel('Density')
plt.show()
No description has been provided for this image
In [57]:
#Bivariate Analysis
#  Box Plot for Car Mileage by Fuel Type
plt.figure(figsize=(8, 6))
sns.boxplot(x='type_of_fuel', y='car_mileage', data=fuel_performance_dataset_clean, palette="Set2")
plt.title('Box Plot: Car Mileage by Fuel Type')
plt.xlabel('Fuel Type')
plt.ylabel('Car Mileage (kmpl)')
plt.show()
No description has been provided for this image
In [58]:
# Violin Plot for Price by Fuel Type
plt.figure(figsize=(8, 6))
sns.violinplot(x='type_of_fuel', y='price', data=fuel_performance_dataset_clean, palette="Set2")
plt.title('Violin Plot: Price by Fuel Type')
plt.xlabel('Fuel Type')
plt.ylabel('Price (INR)')
plt.show()
No description has been provided for this image
In [59]:
# Multivariate Analysis
sns.pairplot(fuel_performance_dataset_clean, hue='type_of_fuel', height=3, diag_kind='kde')
plt.suptitle('Pairplot: Fuel Type, Car Mileage, and Price', y=1.02)
plt.show()
No description has been provided for this image
In [60]:
# FacetGrid: Car Mileage vs Price by Fuel Type
g = sns.FacetGrid(fuel_performance_dataset_clean, col="type_of_fuel", height=5, aspect=1.2)
g.map(sns.scatterplot, "car_mileage", "price", s=100, alpha=0.7)
g.add_legend()
g.set_axis_labels("Car Mileage (kmpl)", "Price (INR)")
g.set_titles(col_template="{col_name} Cars")
plt.suptitle('FacetGrid: Price vs Car Mileage by Fuel Type', y=1.02)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [61]:
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler
import pandas as pd
import matplotlib.pyplot as plt

# Standardize the data (scaling the features so they have unit variance and mean of 0)
scaler = StandardScaler()
df_scaled = scaler.fit_transform(fuel_performance_dataset_clean[['car_mileage', 'price']])

# Apply PCA
pca = PCA(n_components=2)
pca_result = pca.fit_transform(df_scaled)

# Create a color map for fuel type
color_map = fuel_performance_dataset_clean['type_of_fuel'].map({'Petrol': 'blue', 'Diesel': 'green','LPG':'purple','CNG':'red'})

# Plotting the PCA results
plt.figure(figsize=(8, 6))
plt.scatter(pca_result[:, 0], pca_result[:, 1], c=color_map, s=100, alpha=0.7)
plt.title('PCA: Car Mileage and Price by Fuel Type')
plt.xlabel(f'Principal Component 1 ({pca.explained_variance_ratio_[0]*100:.2f}% Variance)')
plt.ylabel(f'Principal Component 2 ({pca.explained_variance_ratio_[1]*100:.2f}% Variance)')
plt.grid(True)
plt.show()

# Print the explained variance ratio
explained_variance = pca.explained_variance_ratio_
print(f"Explained variance by each component: {explained_variance}")
No description has been provided for this image 
Explained variance by each component: [0.59871723 0.40128277]
In [ ]: